The two companies at the front of the AI race are spending more to shape the rules that will govern it. The figures here come directly from the US Senate's lobbying-disclosure database, where lobbying clients must report what they spend each quarter.
A disclosure: Anthropic makes the AI model used in producing this coverage. We report its filings exactly as we report OpenAI's, from the public record.
The numbers
Both companies increased their reported federal lobbying between the first and second quarters of 2026.
OpenAI reported roughly $1.48 million in the second quarter, up from about $1.30 million in the first. The bulk is in-house spending, which rose to $1.2 million from about $1.02 million, alongside steady retainers to three outside firms, Akin Gump, Miller Strategies and DLA Piper.
Anthropic reported roughly $2.87 million in the second quarter, up from about $2.16 million in the first. That is nearly double OpenAI's spend. Its in-house lobbying rose to $1.97 million, and it retains a notably larger stable of outside firms, nine or more in the quarter.
These totals are our own sums of the individual filings each registrant submitted, taken from the line items in the Senate database rather than a headline figure.
The tell is in the issue areas
Dollar totals are less revealing than what the money is aimed at, and here the two diverge.
OpenAI's filings cluster on the expected ground for a frontier-model company: artificial-intelligence legislation and regulation, copyright, data centers and cloud infrastructure, and workforce issues. It is lobbying about the rules for building and deploying models.
Anthropic's roster is broader, and one entry stands out. Its single largest outside retainer in the quarter, about $380,000 to Ballard Partners, is reported against defense, specifically Department of Defense procurement. Defense and national-security themes recur across several of its other firms, alongside export controls and the usual AI-policy and science-and-technology work. In the first quarter that same Ballard retainer was reported at $130,000, so the defense-focused spending roughly tripled quarter on quarter.
That is a meaningful signal. A company weighting a large and growing share of its lobbying toward defense procurement is telling you where it expects a chunk of its revenue, or its strategic importance, to come from: government and national-security customers, not only enterprise and consumer software.
Why it matters for a financial audience
Lobbying disclosures are a cheap, public and surprisingly candid window into corporate strategy, and AI is the sector where the policy stakes are highest.
Three things are worth taking from this quarter's filings.
Both are institutionalizing. Rising in-house lobbying spend, as opposed to just hiring outside firms, is what companies do when Washington becomes a permanent part of the operating model rather than an occasional errand. These are no longer startups treating policy as someone else's problem.
The regulatory fight is being resourced accordingly. AI legislation, copyright, and export controls are live and consequential, and the firms with the most to gain or lose are now spending at a level that reflects it. For investors weighing regulatory risk in AI, the companies' own escalating outlays are a data point about how contested the ground is.
Anthropic's defense tilt is a strategic disclosure. Where OpenAI's spending reads as a company defending and shaping the rules for a commercial product, Anthropic's growing defense-procurement lobbying suggests a company positioning for government and national-security work as a distinct line of business. That is the kind of thing that shows up in a lobbying filing long before it shows up in a revenue segment.
What this is, and is not
The figures are what each company reported it spent, from the official database, and both rose from the first quarter to the second. What the filings do not tell you is outcomes: how effective the spending is, what specific provisions each firm sought, or whether any of it will change a law. Lobbying disclosure captures the effort and its direction, not the result.
But the direction is the point. Two companies building the same core technology are aiming their political money at visibly different targets, and the gap between OpenAI's rules-of-the-road focus and Anthropic's defense lean is a clearer statement of divergent strategy than either company's marketing.
